基于机器学习的地基云图分类研究进展

Research progress of ground based cloud images classification in machine learning

基于机器学习的地基云图分类是光伏发电功率预测的关键技术。该技术主要通过传统机器学习和深度学习方法对地基云图特征提取提升地基云图分类准确率。全文归纳了地基云图分类标准和云图采集设备;简要介绍了地基云图分类数据集;从传统机器学习和深度学习两方面详细论述了典型的地基云图分类方法;比较了不同方法在Kiel F和MGCD地基云图分类数据集上的性能;最后对现有地基云图分类方法进行了总结,并针对目前地基云图分类面临的挑战进行了展望。

Ground based cloud images classification based on machine learning is the key technology of photovoltaic power prediction.This technology mainly improves the accuracy of ground based cloud images classification by extracting features from ground based cloud maps through traditional machine learning and deep learning methods.In this paper,the ground based cloud images classification standard and cloud images collection equipment are summarized,the ground based cloud images classification datasets are briefly introduced,and the typical ground based cloud images classification methods are discussed in detail from two aspects of traditional machine learning and depth learning.Then,the performance of different methods on Kiel F and MGCN ground based cloud images classification datasets is compared.Finally,a summary of existing methods for classification of ground based cloud maps is presented,and an outlook is given for the current challenges of ground based cloud map classification.